DeepMind Chess: Google's AlphaZero AI Beats Champion Software

Demis Hassabis the head of Deep Mind which has already shown it can defeat the world's best Go champions
Credit
JUNG YEON-JE

Demis Hassabis the head of Deep Mind which has already shown it can defeat the world's best Go champions Credit JUNG YEON-JE

The robots are coming for your ... chess game.

Google hasn't commented on the research, but DeepMind's paper said: "Starting from random play, and given no domain knowledge except the game rules, AlphaZero achieved within 24 hours a superhuman level of play in the games of chess and shogi (Japanese chess) as well as Go, and convincingly defeated a world-champion program in each case".

"Starting from random play, and given no domain knowledge except the game rules, AlphaZero achieved within 24 hours a superhuman level of play in the games of chess and shogi (Japanese chess) as well as Go, and convincingly defeated a world-champion program in each case", they said in a paper published by Cornell University in the USA but which has yet to be peer reviewed.

In other words, all of humanity's chess knowledge - and beyond - was absorbed and surpassed by an AI in about as long as it takes to drive from New York City to Washington, DC. AlphaZero won 25 games in which it played with white (gaining the first-move advantage) and a further three in which it played black.

"We now know who our new overlord is", said chess researcher David Kramaley, the CEO of chess science website Chessable. "It will no doubt revolutionize the game, but think about how this could be applied outside chess".

Next came AlphaGo Zero - a fully autonomous AI that learns by teaching itself.

AlphaZero played a 100-game matchup against the world champion, Stockfish 8, and ended up winning or drawing each game. In October, AlphaGo Zero demolished every AI version that preceded it - a victory that an MIT computer scientist told Gizmodo was like watching "an alien civilization inventing its own mathematics".

According to a paper published by the researchers that has yet to be peer reviewed, artificial intelligence took a leap in 1997, when the program Deep Blue defeated the human world champion.

While these feats are impressive in their own right, the true shock is that, unlike DeepMind, AlphaZero was not designed with the ability to master these and similar games in mind.

"I always wondered how it would be if a superior species landed on Earth and showed us how they played chess", grandmaster Peter Nielsen told BBC.

Altre Notizie